© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 8-RES-JB
    11. 10019
    Prev
    Next
    INDEX
    Explanations

    phrases related to different levels or scopes of influence or impact

    oai_token-act-pair · gpt-3.5-turbo

    references to various levels, stages, or aspects of issues and projects

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    onite
    -0.75
    odore
    -0.69
    eka
    -0.66
    agra
    -0.65
     Rost
    -0.65
    odan
    -0.65
    oute
    -0.62
    Dialogue
    -0.60
    password
    -0.59
     Pavel
    -0.57
    POSITIVE LOGITS
     imaginable
    1.57
     conceivable
    1.08
     except
    0.95
     alike
    0.88
     including
    0.85
     whatsoever
    0.82
     ranging
    0.80
     Including
    0.80
     irrespective
    0.80
     regardless
    0.77
    Activations Density 0.205%

    No Known Activations